Voice frequency impact for voice control of collaborative robots

The project is focused on research of voice control of collaborative robots. Previous research of voice control and speech recognition research have given examples and evidence that machines can truly be controlled by human voice. The main problem in the voice control of collaborative robots is the lack of reliability and safety of control compared to manual control. Of the results achieved, the most and least relevant voice control problem has been clearly identified, next step is voice frequency impact for voice control. The experiments were performed in the robotics laboratory at the Faculty of Materials Science and Technology in Trnava, and this using an Arduino kit and a small robot “DOBOT Magician“. In the final part of the project are summarized all experimentally found facts and confronted with the assumptions set in the design of the experiment.


Introduction
Nowadays, we try to make our work easier where possible. The same is true in industry. The robot has become part of many manufacturing processes. In the automotive industry, for example, it has become an integral part of the production process. The growing popularity of robots has brought robotics to households as well. It is no longer a problem to buy a robot for vacuuming, a food processor or a robot controlling the entire smart home. Until recently, voice control was just a dream that people dreamed of. After all, controlling things without moving your hands must be incredibly comfortable. But technology moves big strides forward every day. What was a dream yesterday may become a reality today. Voice control has become part of new cars, smart TVs, smartphones or home helpers. The question is whether voice control has already reached such a level that we can use voice control safely in industry as well. Research into voice control in the past has yielded a lot of new information, such as knowledge about a voice controlled robotic arm vehicle [1], Li-Fi based voice control robot [2], a voice controlled robot using Arduino [3], speech recognition system for a voice controlled robot [4] voicecontrolled autonomous vehicle using IoT [5]. This research again focuses on the issue of voice control safety, uses the knowledge from previous research [6], but also involves the collaborative robot DOBOT Magician. The research is original of its kind, similar research has not yet been carried out.

Objectives and Methods
The aim of this research is to verify the effect of changing the frequency of a voice command on the success probability of voice control of a collaborative robot. For this purpose, the collaborative robot DOBOT Magician and its accessory for voice control, the Arduino kit, were used. DOBOT Magician is a small collaborative robot designed for teaching and simulating various processes, figure 1.  Table 1 shows all voice commands for DOBOT Magician. Previous experience has shown that DOBOT Magician does not have the same success probability for all its voice commands. Voice commands composed of several words (Play music, Close the door) had a significantly higher degree of success probability than voice commands composed of only one word (Up, Go). Therefore, the most reliable voice command (Open the door) was used to determine the effect of changing the frequency on voice control.
The voice command Open the door was programmed so that the robot performed a physical action = movement, which verified whether the voice command was successful.
Sound is any longitudinal mechanical wave in a material environment that is able to evoke auditory sensation in the human ear. The frequency of this wave is approximately in the range of 20 Hz to 20 kHz, outside these limits one does not perceive sound. The human voice handles a fundamental tone of around 400 Hz. Simulations in this research were also performed around the 400Hz spectrum.
The voice command Open the door was recorded with different voices with different frequencies Voice commands were recorded via a headset microphone to a personal computer. The personal computer was equipped with Audacity audio software. The analysis of frequencies of voice commands in Audacity is shown in figure 2. As mentioned, the voice command was recorded in 6 different voice frequencies. In order to determine the effect of changing the voice frequency on the success probability of the voice command, a simulation with 100 identical voice commands with the original frequency was recorded. The results are shown in table 2.

Results
Audacity audio software allows you to change the voice frequency from -99% to + 99%. Simulations in this research were performed at 5% until the DOBOT Magician stopped responding to voice commands.  It is clear from the results that no change in the frequency was able to achieve 100% success probability of voice commands. DOBOT Magician could no longer recognize any of the voice commands when the frequency changed to + 20%.
On the other end, when changing the frequency to -30%, the same applies, none of the voice commands could be recognized by the DOBOT magician. To better represent the results, a table 6 was prepared, which well illustrates the results of this research. The results showed that in all cases, changing the frequency of the voice command contributed to improving the success probability of the voice command. Only in one case, with a voice command with a frequency of 155 Hz, better results occurred after increasing the frequency, in other cases, the improvement was always achieved by decreasing the frequency of the voice command.

Conclusion
This research was to verify the effect of changing the frequency of the voice on the success probability of the voice control of a collaborative robot. Voice commands of different frequencies were selected to achieve the most objective results possible. In 5 of 6 cases, changing the voice frequency by reducing the frequency between 5% and 10% was shown to have a positive effect on the success probability of voice commands. Thus, it is expected in the future that if an input device such as a microphone changes the voice frequency, the success probability of the voice commands will be higher than at the original voice frequency. Voice control can make our work easier and the comfort of using various devices. However, its success probability is still not 100%. The authors of the scientific article still see great scope for improving the voice control of collaborative robots, which can be of great help to us in the future.